Optimal frequency selection for accuracy improvement in binary defocusing fringe projection profilometry

Abstract

The binary defocusing fringe projection profilometry (FPP) technique has demonstrated various advantages for high-speed and high-accuracy three-dimensional (3D) surface measurement. However, higher fringe frequency does not necessarily give better measurements in binary defocusing FPP. To improve the 3D geometry measurement accuracy, this paper proposes an optimal frequency selection approach by analyzing the phase error distribution under different defocusing degrees. The phase error is analyzed theoretically based on the multi-frequency temporal phase unwrapping process, and the associated relationship with fringe frequency, system defocusing degree, noise, and other influencing factors is established. Meanwhile, optimal fringe frequency in a specific system is selected by the theoretical model combined with the validation of simulation experiments. Finally, the measurement accuracy could be effectively enhanced by the generated binary fringe patterns of optimal frequency. Both simulations and experiments verify the effectiveness and robustness of the proposed method.

Adaptive fringe generation with OBB and homography transform for various objects

“…L is the valid pixel l=1,2, …, L along the spindle direction of OBB. It needs to be emphasized that f can be computed by the previous study [18] . Figure 2 showed an adaptive fringe pattern in the FOV of a camera for the bird in figure 1.…”

Fast Three-Dimensional Profilometry with Large Depth of Field

“…By applying 1-bit binary patterns which have much higher projection rates (up to 20 kHz), the binary defocusing technique can greatly improve 3D imaging speed [5][6][7]. The squared binary defocusing method (SBM) is the simplest binarization strategy, which utilizes the binary patterns with the shape of square wave to create sinusoidal fringes [8].…”

“…Hu et al utilized the depth-discrete Fourier series fitting to reduce the complexity of the phase error model [13]. In Zhu's model, more influence factors were taken into account (including defocusing level, intensity noise, and fringe frequency), and the optimal fringe frequency of the binary error-diffusion fringe pattern can be selected [5]. Yu et al achieved an accurate 3D reconstruction in a large DoF by directly transforming the captured patterns into the desired phase with deep learning models [14].…”